In many clinical trials the groups being compared at the end of the trial are not balanced as to initial DMFS scores, even though subjects were randomized to treatment groups initially. Two analytical methods have been used to adjust for these inbalances; the blocking analysis and the covariance analysis. The blocking analysis is the preferred method in that it requires fewer assumptions made of the data. In this study an adaptation of Grainger's severity index, one method of summarizing the initial DMFS scores, was used as a blocking factor in the analysis of caries clinical trial data. This blocking factor has the advantage of being independent of the data set on which it is used. Further, its high correlation with initial DMFS scores insures that it is utilizing the existent information within subjects at the beginning of the trial. This analytical method was compared with 3 other methods, including the covariance analysis and other blocking techniques to see what increase in efficiency would be realized by using this technique.